Full report

Original page 1

NASA Technical Memorandum 104566, Vol. 15 SeaWiFS Technical Report Series Stanford B. Hooker, Editor Goddard Space Flight Center Greenbelt, Maryland Elaine R. Firestone, Technical Editor General Sciences Corporation Laurel, Maryland Volume 15, The Simulated SeaWiFS Data Set, Version 2 Watson W. Gregg Goddard Space Flight Center Greenbelt, Maryland Frederick S. Patt and Robert H. Woodward General Sciences Corporation Laurel, Maryland National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, Maryland 20771 1994

Original page 2

The Simulated SeaWiFS Data Set, Version 2 PREFACE In the prelaunch era of any space mission, simulated data sets are required in order to prove concepts of data formatting; mission health and safety; and mission operations including scheduling, transmission, and capture. Additionally, with a science mission like SeaWiFS, simulated data sets are necessary to test the ground data processing procedures and hopefully, test prelaunch science algorithms. Including the latter functions demands that the simulated data retain a high degree of fidelity to the data expected to be obtained. Producing such a data set for the simulation of scientific observations (including major geophysical variability based on global geophysical fields, sensor physics, and radiative transfer) is time consuming and requires the application of state-of-the-art scientific understanding of the entire remote sensing problem--from sensor physics to the final research algorithms. Version 2 of the SeaWiFS data sets represents a major advance in the generation of a complete simulation. It has already proven extremely useful in testing various components of the SeaWiFS ground processing systems. Hopefully, it will prove valuable to other investigators and institutions who wish to process SeaWiFS data. It will, however, be of limited use for the complete testing of levelo2 and level-3 products with at-launch algorithms. By necessity, it makes use of the preliminary sensor radiometric responses and gain relationships. It is now clear that these will change significantly due to sensor modifications, which are required to minimize stray light effects. Final values of these parameters will be available only after several months from the submission date of this report. Whether there will be a Version 3 of the simulated data sets, including the at-launch sensor responses, has not yet been decided; it will depend upon resources, and how soon before launch the newer version could be ready. Being overtaken by events, in this case sensor modifications, is a risk borne by any simulation system. Greenbelt, Maryland September 1993 -- W. E. Esaias Project Scientist ii

Original page 3

W.W.Gregg,F.S.Patt, andR.H.Woodward ABSTRACT This document describes the second version of the simulated SeaWiFS data set. A realistic simulated data set is essential for mission readiness preparations and can potentially assist in all phases of ground support for a future mission. The second version improves on the first version primarily through additional realism and complexity. This version incorporates a representation of virtually every aspect of the flight mission. Thus, it provides a high fidelity data set for testing several aspects of the ground system, including data acquisition, data processing, data transfers, calibration and validation, quality control, and mission operations. The data set is constructed for a seven-day period, 25-31 March 1994. Specific features of the data set include Global Area Coverage (GAC), recorded Local Area Coverage (LAC), and real-time High Resolution Picture Transmission (HRPT) data for the seven-day period. A realistic orbit, which is propagated using a Brouwer-Lyddane model with drag, is used to simulate orbit positions. The simulated data corresponds to the command schedule based on the orbit for this seven-day period. It includes total (at-satellite) radiances not only for ocean, but for land, clouds, and ice. The simulation also utilizes a high-resolution land-sea mask. It includes the April 1993 SeaWiFS spectral responses and sensor saturation responses. The simulation is formatted according to July 1993 onboard data structures, which include corresponding telemetry (instrument and spacecraft) data. The methods are described and some examples of the output are given. The instrument response functions made available in April 1993 have been used to produce the Version 2 simulated data. These response functions will change as part of the sensor improvements initiated in July-August 1993. i. INTRODUCTION The availability of simulated data is essential for the an improved sensor design and capability, or perhaps a different spacecraft or orbit configuration. In most cases, additional effort and data are required for development of preparation of a remote sensing mission. The usefulness of the simulated data set. the simulated data for mission preparation activities depends upon the thoroughness with which the data are prepared, the adherence to format and content specifications, and the realism of the data. For example, data transfer speeds, storage capabilities, and computer memory requirements can usually be tested using dummy data sets of the expected volume. Unpacking algorithms, computer processing requirements, and intricacies of storing the data require that at least the correct data structures and formats are created in the simulated set. More advanced mission preparation activities, such as quality control, algorithm development, mission health and safety monitoring, and the initiation of an awareness of potential in-flight problems and discoveries, require a realistic data set containing observations similar to the expected flight data, with the correct volumes and structures. Such a simulated data set can potentially increase the success of the mission by exposing problems in sensor design, spacecraft operations, orbit anomalies, etc., well in advance of launch, allowing repair (if detected in time) or compensation by developing algorithms or revising the sensor and spacecraft operations scenario. is attempts to include all information and characteristics to The development of a realistic simulated data set simplified if a predecessor mission exists. If the predecessor mission is an exact copy of the future mission, then development of simulated data is simply a matter of using the previous data. This is rarely the case, however, since the predecessor mission often yields insights that lead to The Sea-viewing Wide Field-of-view Sensor (SeaWiFS) mission, due for launch in 1994, is an example of a mission that has a highly successful predecessor sensor, the NIMBUS-7 Coastal Zone Color Scanner (CZCS), but contains many improvements in sensor design that render direct use of the predecessor data insufficient for mission preparation. The most important differences are the spectral band placement and the global, routine operations scenario for SeaWiFS, since CZCS was a limited duty sensor. Additional differences include the sensor scan parameters and the overall data content, especially the telemetry. This method for constructing a simulated SeaWiFS data set takes advantage of the fact that the eight years of accumulated CZCS data led to many insights into the processes of radiative transfer in the ocean, and the relationship of these processes to remote sensing principles. By emphasizing this knowledge, a realistic simulated data set for a similar, but not identical, sensor may be developed without relying excessively on manipulation of previous, and only approximately valid, data sets. The primary use of this simulated data set is to support SeaWiFS data system testing prior to launch. Thus, it support processing and handling by all components of the SeaWiFS Project. The functions to be tested include, but are not limited to: 1) Transmission from the Wallops Flight Facility (WFF) to the Goddard Space Flight Center's

Original page 4

TheSimulatedSeaWiFSDataSet,Version2 (GSFC)SeaWiFSDataCaptureFacility(DCF); 2) Frameformattingof thedataby the DCF and transferto the SeaWiFSDataProcessingSystem(SDPS); 3) Ingestionby theSDPSandgenerationof levella, level-2,andlevel-3products,includingnavigation,atmosphericcorrection,anddetermination ofgeophysicalparameters; 4) Qualitycontrolby theCalibrationandValidationelement,andnavigationanalysisbytheMissionOperationselement;and 5) Thetransferofdataproductsto theGSFCDistributedActiveArchiveCenter(DAAC),where dataareingested,cataloged,anddistributed. TheVersion2simulateddatareliesuponApril 1993sensor responses.Theseresponsefunctionswill changeaspart of sensorimprovementsinitiatedin July-August1993;therefore,itsvaluefortestinglevel-2andatmosphericcorrection is limited. The methods for constructing the data set are described in the following, along with examples of data output. 2. BACKGROUND The primary scientific goal of the SeaWiFS mission is to provide global monitoring of ocean color. However, because the spacecraft contains limited onboard storage capacity (ll9Mbytes), onboard subsampling of the data is required to obtain global coverage. Thus, SeaWiFS will produce two different resolutions of data: LAC and GAC. LAC data is full sensor resolution (1 km), while GAC data is simply LAC data subsampled every fourth pixel along scan and along-track. GAC data also contains only data within a 45 ° swath width, in contrast to the 58.3 ° swath for LAC data. Since both of these data are stored on board the spacecraft data recorder, they are referred to as stored data. Characteristics of the SeaWiFS sensor and the SeaStar spacecraft are shown in Tables la and lb. Table la. SeaWiFS spectral bands and center Asistypicalin thedevelopmentofsimulateddatasets, wavelengths, shown with those of CZCS for comtheprocessproceedsin steps,eachstepbuildingonthepre= parison. Wavelengths (A) are in nm. viousonein complexityandrealism.Thedatasetversion Sea WiFS CZCS presentedin thispaperis a directdescendantofVersion1 (Gregget al. 1993).ThisversionimprovesonVersion1 in thefollowingways: a) Constructedfora seven-dayperiod,25-31March 1994; b) Includesalldatatypes,i.e.,GlobalAreaCoverage(GAC),recordedLocalAreaCoverage(LAC) andreal-timeHighResolutionPictureTransmission(HRPT)datafortheseven-dayperiod; c) Datacorrespondsto a realisticcommandscheduleforsameseven-dayperiod; d) Includestotal (at-satellite)radiancesnot only forocean,but for land,clouds,andice,andincludesahigh-resolutionland-seamaskbasedon April 1993(linear)responsefunctions; e) IncludespreliminarySeaWiFSspectralrespon- Band No. A Band No. ), 1 412 443 2 443 520 3 490 550 4 510 67O 5 555 750t 6 670 13.5/zmJ; 7 765 8 865 t Intended for surface vegetation analyses only (Williams et al. 1985). Data suspect after 1979 (Williams et al. 1985). In addition to stored LAC and GAC data, SeaWiFS will broadcast real-time LAC data whenever the sensor is turned on. HRPT stations may capture these data whenever the spacecraft is visible. However, due to the agreesesandprelaunchsensorsaturationresponses ment between the National Aeronautics and Space Adprovidedby Hughes/SantaBarbaraResearch Corporation(SBRC),theinstrumentmanufacturer; f) Formattedaccordingto revisedandcurrent(as of this writing)onboarddatastructuresspecifiedbyOrbitalSciencesCorporation(OSC),the spacecraftmanufacturer; g) Includescorrespondingtelemetry(instrument andspacecraft)data;and h) Corresponds to a realistic orbit generated using a Brouwer-Lyddane model with drag. Note that items d) and e) are outdated due to subsequent sensor improvements. ministration (NASA) and OSC, these real-time LAC data will be encrypted: decryption methods will require either a commercial license from OSC, in which case the data may be decrypted in real time, or a research license from GSFC in which case a two-week delay in decryption will be enforced. 3. METHODS An overview of the methods will contribute to clarity and understanding. First, orbit positions are generated for a seven-day period in March 1994, initiated by simulated SeaStar orbital elements. Then a command schedule, which specifies the data collection periods and sensor state,

Original page 5

W.W.Gregg,F.S.Patt, andR.H.Woodward Table lb. Orbitandsensorcharacteristicsforthe SeaWiFSandtheCZCS,shownforcomparison. Characteristic Altitude (km) Period (minutes) Inclination Equator Crossing Time (local) CZCS Sea WiFS 955 705 104.0 98.9 99.28 ° 98.25 ° Noon Noon Node Type Ascending Descending Scan Width Instantaneous Field of View (IFOV) Ground IFOV at Nadir (km) Pixels Along Scan Scan Period (seconds) Scan Plane Tilt Scan Ground Coverage (km) Maximum Spacecraft Zenith Angle Digitization (bits) 39.34 ° 58.3 ° (LAC); 45 ° (GAC) 0.05 ° 0.09 ° 0.825 1.12 1,968 1,285 (LAC); 248 (GAC) 0.134 0.167 (LAC); 0.667 (GAC) =t=20° -t-20 ° 1,566 2,802 (LAC); 1,502 (GAC) 46.8 ° 70.8 ° (LAC); 51.7 ° (GAC) 8 10 Table 2. Simulated NORAD two-line elements for the SeaStar orbit used for orbit propagation in the simulated data set. The important components are: day of year (seventh word, first row), drag term (10th word, first row), inclination (third word, second row), right ascension ascending node (fourth word, second row), eccentricity (fifth word, second row--assume preceding decimal point), argument of perigee (sixth word, second row), mean anomaly (seventh word, second row), mean motion (eighth word, second row), and orbit number (first four digits in ninth word, second row). I LineLine l:O0000Uu93u32uu&j94u80.OOOOOOOOuu.&OOOOOuu98.2330u178.3000uOOlOOOOuuuO.OOOOuuuO.OOOOu14.55400000u29414 is created from these orbit positions based upon a realistic data acquisition requirements scenario, including both routine observations and calibration activities. Then, for each data collection period, the orbit positions to navigate the pointing vectors to Earth locations, according to GAC and LAC sensor specifications and the tilt configuration, is used. The geolocated pixels are associated with a number of atmospheric, land, and oceanic data files to obtain their radiative characteristics. The pixels are determined to be either cloud and ice, land, or ocean, each of which has a different logical pathway for computing at-satellite radiance. Sensor configuration information is appended, such as tilt, gain, etc., along with orbit position and velocity. Telemetry fields are then inserted, using realistic values wherever possible, and dummy fields where' not. Finally, the computed radiances are converted to digital counts and adjusted for the sensor saturation responses using the preliminary linear radiometric response function. The data are stored in two formats: one corresponding to the onboard structure (10-bit science data), and the other corresponding to the ground system frame formatter output (10-bit words in 16 bits, but all other data as in the onboard structure). The frame formatter output is considered to be level-0, since it is the first product seen by the GSFC SOPS and research HRPT stations. The onboard structure is maintained only in the spacecraft flight recorder and at acquisition by the primary ground station, WFF. OOOOO380uuOOOOO-Ouu18855-3uOuu3084 I 3.1 Orbit Propagation Orbit positions and velocities are computed using a modified Brouwer-Lyddane model, with atmospheric drag included. The model is the Simplified General Perturbations Model (SGP4), distributed by the US Space Command. Simulated North American Air Defense Command two-line elements were created for the SeaStar orbit (Table 2) in the so-called NORAD Two-Line Element format, assuming the originally planned launch date of August 1993. The orbit positions and velocities are required for the geolocation step of the simulation, and are also included in the telemetry to simulate Global Positioning System (GPS) data, which are included in the SeaWiFS data. 3.2 Navigation For Version 2, perfect spacecraft attitude control is assumed, and thus roll, pitch, and yaw are set to zero. The tilt configuration involves tilting aft of the velocity vector (toward the North Pole), and changing to fore (toward the South Pole) near the solar declination latitude to minimize sun glint contamination. Geolocation is achieved using an exact algorithm employing vector algebra, developed for SeaWiFS navigation (Patt and Gregg 1993). 3.3 Command Schedules The sensor command schedules used for the simulation correspond to the orbit position data generated for

Original page 6

The Simulated SeaWiFS Data Set, Version 2 the period 25-31 March 1994, and incorporate a routine operations scenario, including GAC observations and calibration activities. The events included in the command schedule are: 1) GAC and LAC record times, including calibration times; 2) real-time HRPT data broadcast times; 3) recorded data downlink times; 4) tilt change times; and 5) sensor configuration for each event. for start up and shutdown, and is included in the schedule. A sample portion of a command schedule is shown in Table 4. Next, individual channel gains and Time Delay and Integration (TDI) settings are selected to maximize the scientific usefulness of the data being collected. For Earthviewing observations, gains are set to unity and TDIs are set to zero, meaning that all four detectors per band are being used to increase the signal-to-noise ratio (SNR). During lunar, solar, intergain, and detector calibration events, gain and TDI settings follow specific instructions (Wood- The process of scheduling begins with knowledge of ward et al. 1993), which are discussed below. The available down-link orbits at WFF. The mission requires two downlinks per day at WFF for full Earth coverage--one near local noon and one near local midnight. The local midnight pass always occurs first in each Greenwich Mean Time (GMT) day as a consequence of the orbit. During flight operations, these will be selected by WFF personnel, who must perform conflict resolution to maximize the acquisition requirements for many spacecraft. For simulation, the orbits are determined by selecting downlink orbits that provide a variety of expected conditions, i.e., as few as six orbits of data collection between downlinks to a maximum of more than nine orbits of data collection (Table 3). The orbit numbers in the table are based on the August 1993 launch date and the late March simulation dates. The selection of downlink orbits to produce variety in data collection schedules simulates a realistic scenario for SeaStar, which is only one of many spacecraft using WFF as the primary data acquisition station. Table 3. Downlink orbit numbers at Wallops used in the command schedule construction. Orbit Number Day of Year 3002 84 3010 84 3017 85 3024 85 3031 86 3038 86 3046 87 3053 87 3060 88 3067 88 3074 89 3083 89 3090 9O 3096 9O Given downlink orbits, the GAC recorder is then scheduled. A 40 minutes-per-orbit maximum duty cycle is adhered to, as determined by SeaStar power constraints. This duty cycle corresponds to a maximum solar zenith angle of 72.7 ° , which is the actual limit used in the scheduling software for GAC data collection. A series of commands is required to configure the instrument and flight recorder 4 gain settings for Version 2 are shown in Table 5. The tilt strategy employed in the command schedules for the simulation data set involves a single tilt change on the descending node. The location of the tilt change is set to minimize sun glint contamination. In flight, the recommendation of Gregg and Patt (1993), which is called the "staggered tilt" strategy, will be followed. This strategy shifts the tilt change location a few degrees north of the maximum sun glint point (near the solar declination latitude) for two days, then shifts the position a few degrees south for the next two days. Gregg and Patt (1993) showed that this strategy improves the ocean coverage over a four-day period while simultaneously reducing sun glint contamination. The SeaWiFS instrument tilt can be commanded to zero (nadir), 20 ° forward and 20 ° aft. For the SeaWiFS convention, an aft 20 ° tilt (opposite the velocity vector) is defined as +20 °, and a forward tilt is defined as -20 ° . Note that the tilt changes are not instantaneous, but require approximately 13 seconds to complete, with tilt change rates that vary continuously during this period (Fig. 1). The sensor tilt position during a tilt change is computed using a mathematical fit to the SBRC data: T = a tanh(bt) (1) where T is the tilt position in degrees, t is the time in seconds and -2O a = tanh(2)' (2) 1 b-- _. (3) The LAC data capacity is determined by the total capacity of the flight recorder and the GAC storage requirements. Calibration and other LAC recording periods must be constrained by the available LAC space. Calibration requirements for the simulation run are derived from a simulated LAC target file. During the mission, this file will be supplied to the Mission Operations team by the Calibration and Validation team. There are five types of calibration activities in the SeaWiFS mission. These are, in order of priority: lunar calibration, solar calibration, intergain check, detector check, and in situ LAC target

Original page 7

W.W. Gregg, F.S. Patt, and R.H. Woodward 2O [] 15 10 O O "o !- 50 [] .o m O O. -5 b- -10 -15 -20 0 1 2 3 4 5 "lime (seconds) [] 6 7 8 9 10 11 12 Fig. 1. Comparison of SBRC measured sensor tilt change response (squares) and mathematical fit used in the simulated data (line). Begin Nominal Sensor Operation t, Sun ...... ....... Sensor Operation Begin Cal Sequence ""-"'-"0 Moon Fig. 2. Lunar calibration maneuver (courtesy of Orbital Sciences Corp.). 5

Original page 8

The Simulated SeaWiFS Table 4. Sample portion of the schedule for day of year 84. Commalld Setting Year L-band Transmitter Turn On 1 1994 Reset Tilt Aft Position 20 1994 Set Earth Mode On 1 1994 Electronics Turn On Power 1 1994 Change Gain Band 1,2,...,8 1 1994 Change TDI Band 1, 2,...,8 0 1994 LAC Transmitting Turn On 1 1994 GAC Recorder Turn On 1 1994 LAC Recorder Turn On 1 1994 LAC Recorder Turn Off 0 1994 LAC Recorder Turn On 1 1994 Change Tilt To Forward -20 1994 LAC Recorder Turn Off 0 1994 GAC Recorder Turn Off 0 1994 LAC Transmitting Turn Off 0 1994 L-band Transmitter Turn Off 0 1994 Electronics Turn Off Power 0 1994 Begin Nominal Sun __.--_..... End Nominal '''"3_5 ----''" Sensor Operation Min Data Set, Version 2 Day T_me Orb/t Latitude Sun 84 r--604 2999 72.32 ° 72.7 ° 84 634 2999 72.32 ° 72.7 ° 84 639 2999 72.32 ° 72.7 ° 84 654 2999 72.32 ° 72.7 ° 84 659 2999 72.32 ° 72.7 ° 84 660 2999 72.32 ° 72.7 ° 84 663 2999 72.32 ° 72.7 ° 84 663 2999 72.32 ° 72.7 ° 84 864 2999 61.27 ° 62.1 ° 84 894 2999 59.22 ° 62.1 ° 84 1680 2999 12.57 ° 13.5 ° 84 1855 2999 1.55 ° 0.4 o 84 1860 2999 0.30 ° 1.7 ° 84 3063 2999 -69.47 ° 72.7 ° 84 3063 2999 -69.47 ° 72.7 ° 84 3065 2999 -69.47 ° 72.7 ° 84 3066 2999 -69.47 ° 72.7 ° Fig. 3. Solar calibration maneuver (courtesy Orbital Sciences Corporation).

Original page 9

W.W.Gregg,F.S.Patt, andR.H.Woodward intergaincheck,detectorcheck,andin situ LAC target observations. For this simulation, the file contains instructions to include one solar calibration per day, which includes an intergain check, and one detector check per day. A full moon occurs during the simulation period, on 27 March 1994, so a lunar calibration event is also scheduled. The file also includes several ship, buoy, and region targets with an accompanying prioritization (Table 6). These requirements are implemented in data collection via the command schedule. Lunar calibration is achieved by pitching the spacecraft 360 ° on the dark side of the orbit to view a moon near full phase (Fig. 2). The SeaWiFS Project has chosen a 7 ° lunar phase angle for lunar calibration (Woodward et al. 1993) to ensure calibration consistency and availability (smaller phase angles do not always occur in a given month). The specific time of the moon view is predicted based on the spacecraft position at the start of the pitch maneuver and the selected pitch rate. The pitch rate has been selected to collect about 20 LAC lines including moon data. The LAC recorder is turned on for two minutes total, centered on the predicted viewing time, to allow for uncertainties in the spacecraft pitch rate. Specific gain settings are utilized to prevent saturation and to ensure maximum dynamic range (Table 5). Table 5. Gain settings for SeaWiFS. Band Gain 1 Gain 2 Gain 3 Gain 4 412 1 2 3 1.7 3 1.3 443 1 2 1.73 1.3 490 1 2 1.7 1.25 510 11 2 1.7 1.252 555 12 2 1.61 0.77 670 1 2 0.71 0.462 765 1 2 0.611 0.322 865 1 2 0.551 0.262 1. Used for lunar calibration. 2. Used for solar calibration. 3. Used for both lunar and solar calibration. of and oceans at the time of observation. This is obtained Solar calibration involves tilting the sensor 20 ° aft the velocity vector when the spacecraft is at the southern terminator to view reflection off a solar diffuser plate mounted on the sensor (Fig. 3). The instrument is commanded into a unique solar mode configuration, which causes the sensor to sample a different part of the scan circle to allow a view of the diffuser plate, although, the LAC scan line still consists of 1,285 science pixels. Again, specific gain settings are used to maximize the scientific usefulness of the solar irradiance scattered off the diffuser plate (Table 5). The solar mode also includes an internal calibration signal. During each scan, after viewing the diffuser plate, the sensor electronics receives a constantmagnitude electronic pulse, called the calibration pulse, which enables a highly consistent intergain check to be performed. This check is performed immediately after the full solar calibration, using a specific sequence of channel gains (Table 7). Detector checks also use the solar diffuser plate, and are performed on the orbit following the solar calibration activity. Again, a specific TDI and gain sequence is used (Table 8). Finally, the remaining LAC recorder space is used to schedule in situ calibration targets. A distinction is made between two types of in situ targets for the SeaWiFS mission: ship and buoy targets, and region targets. Ship and buoy targets are discrete locations which have the higher priority of the two in situ types. For ship and buoy targets, the LAC recorder is turned on 15 seconds before viewing the target, and left on for 15 seconds after the target. This 30 second viewing duration provides 100 km of observations about the target. For regions, the recorder is turned on only if the sub-satellite point is within the region. This ensures that at least half the scan is in the region of interest. In this simulated data set, not all the available LAC space is used, due to the prototype status of the scheduling algorithms and code, which are under development and testing. For the mission, all available LAC space will be used. 3.4 Determining At-Satellite Radiances At-satellite radiances are derived from five backscattering sources: clouds, land, ice, non-cloud atmosphere, and the ocean. In each case, the method for computing radiances involves the assumed known reflectance properties of the contributing source; solar geometry, which is dependent on the year, time of day, and position; and scan geometry, which is dependent on the simulated orbit position and known scanning characteristics of the sensor. Intensive radiative transfer calculations translate this knowledge into quantitative radiances taking into account the specific optical responses of the SeaWiFS sensor available as of April 1993. The calculations have not yet been made for a bilinear instrument response. The calculation of at-satellite radiances thus begins with a determination of the state of the atmosphere, land, through the use of several global data sets. The same seven-day period in March 1990 for each data set was chosen to ensure coherence among the state variables. The data sets include information on cloud cover, ice cover, land vegetation, ocean chlorophyll concentration, wind speeds, surface pressure, and ozone concentration (Table 9). Each of these variables affects SeaWiFS observations: 1) Cloud cover determines the ability of the sensor to view the Earth and also determines the amount of saturation; 2) Ice cover and land vegetation affect the optical properties of land; 3) Ocean chlorophyll concentration contributes to the determination of the spectral properties of water-leaving radiance; 7

Original page 10

The Simulated SeaWiFS Data Set, Version 2 Table 6. LAC target file used for the simulated SeaWiFS data set. In situ targets have longitude and latitude associated with them, plus a priority indicator. Calibration activities are associated with a number indicating the number of such activities per month. Calibration Targets 1994 82 In situ i0 Clark' s Buoy -156.3400 Bermuda Buoy -71.9000 JGOFS 63.2500 NOAA S. Atlantic Bight -77.5200 NOAA Gulf of Cal. -107.2600 18.6700 4 32.1200 6 19.4000 3 32.0300 4 22.1100 5 S. Africa 10.2300 -32.8700 6 Galapagos -92.6200 Gulf of Mexico -86.8600 Oregon St. -131.1400 Navy Bering Sea -175.5800 Regions 5 -3.2500 8 24.7900 3 45.7700 1 63.4200 10 Sargasso Sea -70.0000 -45.0000 20.0000 30.0000 2 Gulf of Mexico -110.0000 -80.0000 17.0000 31.0000 1 SE Pacific -150.0000 -90.0000 -60.0000 -30.0000 3 Galapagos -105.0000 -75.0000 -15.0000 0.0000 5 Micronesla 135.0000 180.0000 0.0000 15.0000 4 Solar Calibration 30 Lunar Calibration 1 Intergain Calibration 30 TDI Check 30 4) Wind speeds determine the aerosol type over the oceans and the sun glint radiance magnitudes; 5) Surface pressure affects the Rayleigh scattering of the atmosphere and the oxygen concentration which absorbs irradiance and radiance in some SeaWiFS bands; and 6) Ozone concentrations affect the spectral transmittance of irradiance and radiance through the atmosphere. The effect of water vapor on SeaWiFS bands is small, and so is set to a global mean of 1.5 cm. As stated above, the simulation procedure includes orbit propagation and navigation to determine the position of each pixel. The of enforces a Lambertian scattering assumption. For both pixel's position is associated with the known position each variable in the data sets, enabling knowledge of the current state. Examples of cloud and ice cover used are shown in Figs. 4 and 5. Given knowledge of the state of the atmosphere and surface, and the pixel under examination in a full orbit propagation, the radiative properties are determined next. For cloud and ice cover, concentrations greater than 50% means that obscuration of the surface is assumed and the at-satellite radiance is computed as: Lt(A) = pc,iEd(A, Oo), (4) 7r where Lt is the total radiance received by the satellite (atsatellite radiance), p is the reflectance of clouds (subscript c) and ice (subscript i), and Ed is the incident irradiance on the Earth's surface as a function of the Earth-sun distance and atmospheric transmittance. The atmospheric transmittance properties are determined by the solar zenith angle, 00, which is known at each pixel from orbit position, pointing knowledge, and geolocation. Division by lr clouds and ice, p is set to 1, i.e., clouds and ice are completely reflecting. The incident irradiance is computed using the method of Gregg and Carder (1990), where extraterrestrial irradiance and other atmospheric optical properties are weighted to conform to the latest SeaWiFS spectral responses (Table 10). If a pixel is found to be greater than 50% ice or

Original page 11

W.W. Gregg, F.S. Patt, and R.H. Woodward 4_ o_ _9 _9 O _D _D CI2 g,q o ¢D o 9 a o g

Original page 12

The Simulated SeaWiFS Data Set, Version 2 _9 9 {D ° _9 O L) 9 C$3 b-4 c9 ¢9 ° 10

Original page 13

W.W. Gregg, F.S. Patt, and R.H. Woodward cloud, (4) is utilized and no further calculations are performed, under the assumption that the high reflectance properties of ice and clouds overwhelm other atmospheric, non-ice land, and oceanic contributions. If a pixel is found to be non-cloudy and non-ice, then it must be determined whether the pixel is land or ocean. This determination is made using a high resolution land-sea mask, which is derived from the World Vector Shoreline (WVS) Database, available from the National Geophysical Data Center (NGDC). The resolution of the land or sea mask is 128 pixels per degree of latitude or longitude, corresponding to 0.9 km at the equator, or slightly better than the LAC resolution. If the pixel under observation is found to be land, then the land radiance determination method is employed. Otherwise, the ocean radiance determination method is used. Table 7. Command sequence for intergain check. Gain 1, Gain 2, Gain 1, Gain 3, Gain 1, Gain 4, Gain 1 This gain sequence is repeated for each of 5 TDI configurations: 1. Use of all four detectors 2. Use of Detector 1 only 3. Use of Detector 2 only 4. Use of Detector 3 only 5. Use of Detector 4 only The overall pattern is repeated once, along with a final repeat using all four detectors, requiring a total of 77 seconds to complete. Table 8. Command sequence for a TDI check. This sequence is performed using solar calibration gains on all bands, since the check occurs in solar mode using the solar diffuser plate. Unique TDI configuration codes are defined by OSC, and these are used in the simulated data set. TDI Configuration Code Meaning 0 All 4 detectors 164 Detector 1 only 0 26 Detector 2 only 0 74 Detector 3 only 0 161 Detector 4 only 0 The pattern is repeated 9 times, requiring 81 seconds The land radiance determination method begins with the Advanced Very High Resolution Radiometer (AVHRR) reflectance data set. First, it is assumed that the reflectance of AVHRR band 1 (centered at 640 nm) equals the reflectance at SeaWiFS band 6 (670 nm). The Normalized Difference Vegetation Index (NDVI) is then computed. If the NDVI is greater than 0.2, the pixel is vegetated, and it is assumed that the reflectance at AVHRR band 2 (centered at 850 nm) equals the reflectances at SeaWiFS bands 7 and 8 (765 and 865 nm, respectively). Table 9. External data sets used in the creation of simulated SeaWiFS data and their sources. External Data Set Source Chlorophyll Global CZCS composite (CZCS Global Reprocessing) Land Vegetation 10-day AVHRR composite from April 1988 (AVHRR Pathfinder) Wind Speed and FNOC data from Surface Pressure 25-31 March 1990 (NCDS) Clouds ISCCP from 25-31 March 1990 (NCDS) Ice ISCCP climatological mean (NCDS) Ozone TOMS from 25-31 March 1990 Table 10. Spectrally weighted (FWHM) mean extraterrestrial irradiance F0(A) (mW cm -2/_m-1), Rayleigh optical thickness (vr), ozone absorption coefficient (aoz), water vapor absorption coefficient (awv), and oxygen absorption coefficient (ao). All units are in cm -1 , except Tr, which is dimensionless. A F0 Tr aoz awv ao 410 _71.07 0.3139 0.0000 0.0000 0.0000 443 189.91 0.2341 0.0030 0.0000 0.0000 490 194.33 0.1561 0.0213 0.0000 0.0000 510 188.24 0.1324 0.0382 0.0000 0.0000 555 185.93 0.0942 0.0897 0.0008 0.0000 670 152.14 0.0439 0.0470 0.0043 0.0053 765 123.55 0.0257 0.0083 0.0007 3.3780 865 100.00 0.0155 0.0000 0.0057 0.0000 Reflectances are then computed at SeaWiFS bands 1- 5 as an inverse function of chlorophyll absorption coefficients weighted by the SeaWiFS spectral response, where the magnitude conforms to the reflectance at band 6. Values for chlorophyll absorption are taken from Gregg et al. (1993). If the NDVI is less than 0.2, the pixel is nonvegetated, and the slope is computed from AVHRR band 2 minus AVHRR band 1. A linear reflectance response, through SeaWiFS bands 1-5, is assumed following this slope. This method provides reasonable agreement with observations for two arbitrarily selected vegetated and nonvegetated points (Fig. 6). The reflectances are converted to at-satellite radiances as in (4), with the reflectance determined above substituted for Pc,i- Note that this method does not explicitly account for the atmosphere, although some is implicit in the AVHRR reflectances which do not correct for the atmosphere, nor for scan angle. The ocean radiance method is described in great detail in Gregg et al. (1993) and will not be elaborated upon in 11

Original page 14

The Simulated SeaWiFS a) 50-- Green Vegetation 40- 30-I Dry Soil 2o-I (an_ewS:il, __ rr 10- 400 500 600 b) 20- 15- . 5- I I 400 500 600 Data Set, Version 2 / .--'"'.=-=['_ 700 800 900 1000 Vegetation I I I 700 800 900 1000 Wavelength (nrn) Fig. 6. a) Land reflectance data from Tucker and Miller (1978). b) Simulated land reflectances from AVHRR data for a vegetated location near the Nicaraguan and Costa Rican border, and for a nonvegetated location in the Mexican plateau (redrawn i 12 i with permission from J. Tucker).

Original page 15

W.W. Gregg, F.S. Patt, and R.H. Woodward this document. There are, however, three significant differences in the method used here. First, wind speeds are taken from Fleet Numerical Oceanographic Center (FNOC) data for 25-31 March 1990, rather than using the climatological mean as in Gregg et al. (1993). Second, pressure effects are explicitly accounted for in the computation of multiple Rayleigh scattering through 1 -- e -r'(X)M(O) I'(A) = Io(A)__- ), (5) where I0(A) is the Rayleigh scattering intensity at standard pressure (Gordon et al. 1988), fro is the Rayleigh optical thickness weighted by the SeaWiFS spectral response, and M is the slant path length through the atmosphere (see Gregg et al. 1993). The term T is the pressure-corrected optical thickness P (6) sensor exhibited significant saturation response. This re- = 0' where P and P0 are the local surface pressure and the standard pressure, respectively. Finally, the factor 4_ is eliminated in the denominator of (28) in Gregg et al. (1993), since it is already accounted for in (30) of that reference. This error produced incorrectly low aerosol radiances in Version 1. 3.5 Conversion to Digital Counts SeaWiFS simulated total radiance data were converted into digital counts by DC(A, G) = 20 + Lt(A) - LNER(A, G) Table 11. LNER and saturation radiances (Lsat) for SeaWiFS (units are mW cm -2 #m -1 sr-X). Lsat will change in the launch configuration due to instrument modifications. Band A [nm] LNER Lsat 412 0.0094 13.63 443 0.0085 13.25 490 0.0055 10.50 510 0.0050 9.08 555 0.0041 7.44 670 0.0037 4.20 765 0.0022 3.00 865 0.0015 2.13 3.6 Sensor Saturation Response The preliminary April 1993 test data from the SeaWiFS sponse has two main features: 1) An optical effect due to stray light in the sensor that produces elevated digital counts preceding a saturated pixel, and 2) An electronic effect, e.g., bright target recovery (BTR), that produces elevated counts following a saturated pixel. Furthermore, these effects are asymmetric with band number; even numbered bands exhibit greater stray light effects, while odd numbered bands exhibit greater bright target recovery effects. These effects were added to the simulated data to enhance realism. Using data from a sensor characterization examination in April 1993, methods G) ' (7) were developed to simulate the stray light and BTR rewhere DC is the digital count value, LNER(A,G) is the Noise Equivalent Radiance (NER) as a function of wavelength and gain factor G, and s is the slope for the range 0-1,023. This is applicable for 10-bit digitization, given by s(A,G) -- Lsat(A,G) - LNER(A, G) 1003 , (8) where Lsat(A, G) is the saturation radiance for the sensor as a function of wavelength and gain setting. LNER()% G) and Lsat(A,G) for Gain 1 are shown in Table 11. The addition of 20 counts in (7) represents the dark restore value used in the simulation. In flight, a dark restore value will be contained in the data stream and will be applicable for each scan line. This dark restore value also accounts for the division by 1,003 in (7), rather than division by 1,023. Application of (7) and (8) does not truncate the data to 10-bit quantization. In the first run of the simulation of These fields represent a combination of direct measuredata set, this is desirable, since it provides knowledge the extent of saturation (number of counts above 1,023), which is required for determination of the sensor saturation responses. After the saturation responses are accounted sponses. Both are a function of the level of saturation, i.e., the real digital count value for a sensor not limited to 10-bit words. In the calculation discussed earlier, count values were saved as computed, i.e., not truncated to 10-bit accuracy as is the case for the actual flight. The sensor characterization data provided the saturation responses at two levels of saturation, or overdrive, for each pixel in the scan line. A linear function was fit individually for each pixel away from the saturated pixel. This linear function fit provided an excellent reproduction of the tabulated saturation responses (Figs. 7 and 8). 3.7 Telemetry The SeaStar data stream contains a large number of instrument and spacecraft telemetry fields in addition to the SeaWiFS science observations contained in the scan lines. ments from spacecraft hardware components, e.g., temperatures, voltages, and currents; status of commandable components, e.g., on or off, and A/B side switches, as well for, the final simulation data set values are truncated to a as tilt and gain settings; and computed values from onmaximum count value of 1,023. board processors, e.g., spacecraft time, orbit position, and 13

Original page 16

TheSimulatedSeaWiFSDataSet,Version2 -2 o _unoo 14 ° t , o O Z f!i ° _unoo

Original page 17

W.W. Gregg, F.S. Patt, and R.H. Woodward t,=, oq:: 0 o_ . @ ,,,,_ o e_ _unoo _,unoo 15

Original page 18

The Simulated SeaWiFS attitude. Fields that represent physical quantities (both measured and computed), also known as analog telemetry, are usually digitized for inclusion in the data stream, while status information, also known as discrete, is typically represented by single bit fields packed into status bits. Several telemetry fields are required as input to the simulation process. They are: a) scan line time, b) orbit position and velocity, c) spacecraft attitude, d) instrument tilt angle (commanded and actual), e) gain and TDI settings, and f) instrument mode. A number of additional fields are also included to support data processing and otherwise improve the realism of the data stream. Specific examples include instrument analog and digital telemetry, spacecraft attitude sensor data, and GPS telemetry. In addition, the spacecraft telemetry includes a large number of analog and status fields, from the attitude control system data, GPS subsystem, electrical power subsystem, propulsion subsystem, transmitters, and flight computers. In most cases, the inforis of 10-bit and 8-bit fields. All data, except the spacecraft mation presently available for simulating these values limited. The methods for each type of data are as follows. The simulated science data discussed earlier are saved scan line-by-scan line. The scan line time and instrument mode are inserted for each LAC scan line and for every fifth GAC scan line, since GAC lines are stored five per minor frame (discussed in the next section). The gain and TDI are included with every scan line. Time, sensor tilt, and mode are required for the instrument telemetry packet, included for two of every three scan lines for both GAC and LAC. Time, orbit position and velocity, and spacecraft attitude are required for the ancillary data packet, included for one of every three scan lines. The remaining instrument telemetry fields are set to static values based on a sample of actual values provided by SBRC. The spacecraft telemetry packet is included for 1 of every 3 scan lines for LAC data and 1 of 15 lines for GAC data, in accordance with the OSC specified onboard data structure. The scan line time, orbit, and attitude data for the appropriate scan line are included in the spacecraft telemetry. In addition, simulated attitude sensor data is generated for the spacecraft packet to support realistic navigation processing by the SDPS. The sensor data consist of sun angles and presence flags for each of three digital Data Set, Version 2 3.8 Data Formatting The data is formatted according to the specifications provided by OSC for the SeaStar minor frame and the individual data packets. The LAC format is described in the SeaStar Spacecraft L-band Downlink to Receiving Stations Interface Control Document (ICD), currently being prepared by OSC. This document includes the overall minor frame format and details of the spacecraft identification (ID) and time tag, scan line data (including the gain and TDI), instrument telemetry, and ancillary data packets. However, the structure of the data fields is shown here for completeness (Fig. 9). The GAC minor frame format specifications were obtained from OSC, but are not yet available in a formal document. The formats of the GAC data ID, time tag, spacecraft telemetry packet, instrument telemetry packet and ancillary data packet are identical to the LAC data, although, the instrument and ancillary packets are stored in different locations in the GAC minor frame (Fig. 10). The GAC scan lines are formatted and stored according to the data reduction specifications provided by OSC. The overall format of the LAC minor frame is given in Fig. 9, which shows the location of the major elements. As stored on the spacecraft, the data include a combination telemetry, are stored as 10-bit words; the spacecraft fields are 8-bit aligned. The figure also shows a frame synch at the beginning of each minor frame and an auxiliary synch at the end; the simulation of these fields was not attempted, but will be added by the DCF to support endto-end testing. The Version 2 format is based upon this minor frame definition and on the initial frame formatting of the data. Clearly the 10-bit aligned data cannot be readily interpreted by standard computing systems. The DCF has provided a programmable frame formatter (PFF) which reformats the data into a more useful format. Specifically, each 10-bit word is right-justified in a 16-bit word, while the 8-bit words are maintained in their original format. The frame synchs are stripped and discarded. This processing increases the frame length from 13,860 bytes to 21,504 bytes, but greatly facilitates subsequent processing. The data in the format generated by the PFF is referred to as level-0, since it is the first format seen by the SDPS or any remote HRPT station. The simulation reverses the processing performed by the DCF. First the simulated science and telemetry data are combined to generate a PFF (level-0) output file, with all data aligned in 16- or 8-bit fields. This format is not sun sensors, and Earth width and phase angles for each of only much easier to generate initially, but also facilitates two Earth scanners. The angles are computed using the simulated orbit and attitude data, according to the sensor manufacturer's specifications and information from OSC regarding the sensor mounting and orientation, as well as general reference information (Wertz 1978). The remaining spacecraft fields are set to arbitrary static values. 16 reading of the data by other software for quality control (QC); it can be ingested and processed directly by the SDPS for internal testing of its systems and all downstream functions (QC and data archiving). The level-0 files are then processed to compress the 16-bit words to 10 bits to simulate the spacecraft format.

Original page 19

W.W.Gregg,F.S.Patt, andR.H.Woodward v _ < ,. . ig ?_ .- ._ o_o#,o#o o o# 14= ",_ <._= =_ o .......... E a : IxQ . _ io,,, o _ i_ i_ ! i_ e- LL _ o _ _ v 6 {D L. 1 1 I + t..) 1 1 v * ,_ _ _ 0 '_ ® E _ ,.a - ,,. 0 M) O 0 0 z • I?

Original page 20

The Simulated SeaWiFS ul "o e- 0 o 4) (n >, lid G) :> G) E o ,m E c o u O =, > II o ,m x_ t- "[3 o o .i L_. ¢,1 o E 4) Lq_ O <f o .J O O 5 < < < 18 Data Set, Version 2 o L) U U -a 4_ 0 12 o L) .£ q= b $ o (D

Original page 21

W.W. Gregg, F.S. Patt, and R.H. Woodward Table 12. SeaStar spacecraft telemetry engineering conversion of analog telemetry. Analog Telemetry Points Units Start Byte Orbit X position (ECEF) m 143 Orbit Y position (ECEF) m 147 Orbit Z position (ECEF) m 151 Orbit X velocity (ECEF) ms-I 155 Orbit Y velocity (ECEF) ms-I 159 Orbit Z velocity (ECEF) ms-i 163 Attitude yaw angle degrees 121 Attitude roll angle degrees 123 Attitude pitch angle degrees 125 Sun sensor 1 angle 1 degrees 61 Sun sensor 1 angle 2 degrees 63 Sun sensor 2 angle 1 degrees 67 Sun sensor 2 angle 2 degrees 69 Sun sensor 3 angle 1 degrees 73 Sun sensor 3 angle 2 degrees 75 Earth scanner 1 phase degrees 79 Earth scanner 1 width degrees 81 Earth scanner 2 phase degrees 85 Earth scanner 2 width degrees 87 ECEF = Earth-Centered Earth-Fixed Place holders are inserted for the frame synch words to be filled by the DCF. These files are then used by the DCF and the GSFC Code 500 Simulation Operations Center (SOC) to generate simulated data tapes which can be used to test the WFF data receipt, transmission to GSFC, and PFF processing. A description of the DCF and SOC sim- Byte Slope Intercept Length 0.00391155 -84OO0OO.O 0.00391155 -84OOOOO.0 0.00391155 -8400000.0 4 3.49246E-6 -7500.0 4 3.49246E-6 -7500.0 4 3.49246E-6 -7500.0 2 0.00549316 -180.0 2 0.00549316 -180.0 2 0.00549316 -180.0 2 0.0O1953125 -64.0 2 0.001953125 -64.0 2 0.001953125 -64.0 2 O.001953125 -64.0 2 0.001953125 -64.0 2 0.001953125 -64.0 2 0.005493164 0.0 2 O.005493164 0.0 2 0.005493164 0.0 2 0.005493164 0.0 and are therefore, not useful for data processing. All fields are digitized using a linear conversion (with a specified slope and offset) and stored in either 16 or 32 bits. The remaining spacecraft packet bytes are set to arbitrary values (determined as the byte offset modulo 256). The instrument telemetry specifications are shown in ulation support is beyond the scope of this paper. All Table 13. Note that, although the instrument data are of the scan lines and telemetry corresponding to each data collection period are collected and formatted as a unit. The methods involved in formatting the individual fields and packets for the minor frames are different for each type of data and are as follows. The spacecraft ID is computed for each minor frame from the frame number and the data type. As shown in Fig. 9, there are three minor frames per major frame; the frame number is started at 1 at the beginning of each data collection period and cycled for all frames in the period. The time tag is converted to two integer fields corresponding to the truncated Julian day (days since 13 January 1993, at midnight GMT, Julian day 2,449,000.5) and the milliseconds of the day; the significant bits of these two fields are combined into a 40-bit time tag, which is divided into four 10-bit words. The spacecraft telemetry fields are digitized and stored according to the best available specifications from OSC (Table 12). Note that these specifications have not been updated since September 1992 and are very likely to change prior to launch. Spacecraft fields are included only for the first minor frame of each major frame, corresponding to included in the 10-bit portion of the minor framel in fact only the 8 least significant bits (LSB) of each word are used. The instrument analog telemetry fields also use linear digitization specifications. The measured tilt angle is the only actively simulated analog value. The tilt alignment status and the sensor mode (Earth or solar) are used to set the appropriate discrete status bits; the mirror side bit is alternated for the LAC scan lines. The time tag is set to the milliseconds since the last exact second. The remaining fields are set to fixed, but realistic, values as stated above. The ancillary data specifications are shown in Table 14. The ancillary data fields use signed integers rather than linear conversions for digitization; the units are meters and meters per second for the orbit data, and microradians for the attitude angles. The time tag is set to the millisecondsof-day. The gain and TDI for each band are combined into a single 10-bit field, with the 8 fields corresponding to the 8 bands immediately preceding each scan line in the minor frame, as shown in Fig. 9. (Note that the gain values in spacecraft (S/C) state of health (SOH) telemetry, (field 1 the data are 0 through 3, corresponding to gains 1 through in Fig. 9). OSC has designated fields 2 and 3 as dynamic 4 in Table 5,) Each scan line is also preceded in the frame 19

Original page 22

The Simulated SeaWiFS Table 13. SeaStar instrument telemetry specification. Word Bit Time Stamp of Sync Pulse Number Number in Scan Line Data Set, Version 2 Word SeaWiFS Analog Number Telemetry Milliseconds Since Last GPS Pulse 8 Band 1 and 2 FPA Temperature 2 Milliseconds Since Last GPS Pulse 9 Band 3 and 4 FPA Temperature Milliseconds Since Last GPS Pulse 10 Band 5 and 6 FPA Temperature Milliseconds Since Last GPS Pulse 11 Band 7 and 8 FPA Temperature Word No. Bit No. SeaWiFS Discrete Telemetry 5 1 Pad 5 2 Pad 5 3 Servo A or B Select 5 4 AMC On 1 5 5 Servo A Locked 5 6 Servo B Locked 5 7 Timing A or B Select 5 8 Tilt A On 5 9 Tilt B On 5 10 Tilt Telemetry On 6 1 Pad 6 2 Pad 6 3 Stow On 6 4 Stow Aligned 6 5 Heaters Enable 6 6 Solar Door Status 6 7 Analog Power On 6 8 Tilt Platform Limit 6 9 Tilt Base Limit 6 10 Tilt Nadir (0 °) Aligned 7 I Pad 7 2 Pad 7 3 Tilt Aft (+20 ° ) Aligned 7 4 Tilt Forward (-20 °) Aligned 7 5 Data Mode Select 7 6 Half Angle Mirror Side 7 7 Image Data Sync 7 8 AMC at Speed 7 9 AMC at Speed 7 10 Spare AMC = Angular Momentum Compensation FPA = Focal Point Assembly a tion of the sensor saturation response. The final step prior by start synch and dark restore pixels, and followed by stop synch pixel. Thus, the total number of stored pixels per scan line is three more than the actual number of data pixels (1,288 for LAC and 251 for GAC). The start and stop synchs are set to alternating values of 0 and 1,023 (10 binary ones) based on information provided by OSC, while the dark restore value is arbitrarily set to 20 for all bands. The scan line radiances are initially computed using the full 16-bit dynamic range, to allow realistic computa- 2O 12 Telescope Motor Temperature 13 Tilt Base Temperature 14 Tilt Platform Temperature 15 Half Angle Motor Temperature 16 Power Supply A Input Current 17 Power Supply B Input Current 18 +15 V Analog Power Voltage 19 -15 V Analog Power Voltage 20 +5 V Logic Power Voltage 21 Power Supply Temperature 22 B1 and B2 Post-amplifier Temperature 23 Servo Driver Temperature 24 +30 V Servo Power Voltage 25 +21 V Servo Power Voltage 26 -21 V Servo Power Voltage 27 +5 V Servo Power Voltage 28 AMC Phase Error 29 Tilt Platform Position 30 Tilt Base Position 31 +28V Heater Power 32 Telescope A Motor Current 33 Telescope B Motor Current 34 Half-Angle Mirror A Motor Current 35 Half-Angle Mirror B Motor Current 36 Servo A Phase Error 37 Servo B Phase Error 38 AMC A Motor Current 39 AMC B Motor Current Word No_ Padding to Fill 44 Word Field 40 Spare 41 Spare 42 Spare 43 Spare 44 Spare to inclusion in the minor frame is to truncate all values to 1,023 (maximum 10-bit value). The completed minor frame is then written to the level-0 simulated data file. 4. EXAMPLES OF DATA This section includes examples of GAC, stored LAC, and real-time HRPT data.

Original page 23

W.W. Gregg, F.S. Patt, and R.H. Woodward 4.1 GAC Data Simulated Version 2 data for SeaWiFS band 1 for the second (daytime) downlink on 25 March 1994, mapped to Earth coordinates, shows how the descending node orbit acquires GAC data (Fig. 11). This image shows eight full descending nodes of data, plus a partial node over Wallops prior to a daytime downlink (the Wallops visibility mask is illustrated). Approximately half of an overpass at Wallops results in downlinked data for that opportunity since the spacecraft contains two onboard data recorders--one transmits while the other acquires. The remaining half of the data taken during the overpass will be received at the next (night) downlink. In Fig. 11, the grey scale shows lower (darker) to higher (lighter) count values representing total, at-satellite radiances. White indicates saturation; in this case, due almost entirely to clouds and ice. (See also Plate 1, where a blueto-green-to-red scale is used, and the brightest red indicates saturation.) Some saturated pixels, however, may be noted in the Saharan Desert. Note that even clouds and ice do not saturate at higher latitudes due to reduced incident irradiance. The pinching of the swath width near the equator shows the effects of the tilt change. The tilt change requires the sensor to pass through nadir, where the swath width is reduced relative to tilted scans. The black speckles in this region indicate lost coverage due to excessively large effective ground speed, a function of the orbital speed and the tilt change speed (see Fig. 1). It is not clear how well data collected during a tilt change will be navigated due to attitude perturbations resulting from the tilt motor. In this simulated data set, no such perturbations were allowed to occur and pointing information is as well known as anywhere in the orbit. Other than navigation difficulties and a large occurrence of sun glint, data collected during a tilt change is expected to be of high scientific quality. Whereas band 1 saturates only over clouds, ice, and parts of the Saharan Desert, band 8 saturates over nearly all land features (Fig. 12; Plate 1). This is because vegetated areas have large reflectance in the near-infrared regions of the solar spectrum (Fig. 6). A relatively cloud-free portion of a GAC orbit was selected to show how GAC data will actually look, as received, in the data stream. A section for band 1 shows Haiti and the Dominican Republic--near the center of the image--as dark, with Puerto Rico slightly to the east (Fig. 13; Plate 2). Greater variability is seen over the land than over the oceans. This is because ocean at-satellite radiances are dominated by the atmosphere (approximately 90% of the total), and in particular, Rayleigh scattering. Some minor brightening may be seen near the scan edges due to increased Rayleigh scattering as a function of the greater atmospheric path length here. An image of band 8 shows land as saturated once again (Fig. 14; Plate 2). The blocky, rectangular features now Table 14. SeaStar ancillary telemetry specification. Word No. Time Reference 1 Milliseconds since Midnight UTC (MSB first) Milliseconds since Midnight UTC Milliseconds since Midnight UTC Milliseconds since Midnight UTC (LSB last) Word No. Orbit Position Data 1 5 X (MSB first) 2 6 X 7 X 8 X (LSB last) 9 Y (MSB first) 10 Y 11 Y 12 Y (LSB last) 13 Z (MSB first) 14 Z 15 Z 16 Z (LSB last) Word No. Orbit Velocity Data 3 17 (MSB first) 18 )( (LSB last) 19 (MSB first) 20 1I (LSB last) 21 ,_ (MSB first) 22 2 (LSB last) Word No. Attitude Angular Position Data 4 23 (I) (MSB first) 24 (I) (LSB last) 25 0 (MSB first) 26 0 (LSB last) 27 (MSB first) 28 (LSB last) Word No. Attitude Angular Rates Data 5 29 (MSB first) 30 (LSB last) 31 ) (MSB first) 32 (LSB last) 33 t (MSB first) 34 (LSB last) 35-44 Spare 1. ECEF Cartesian, units are in meters. 2. MSB is the acronym for most significant bits. 3. ECEF Cartesian, units are in ms -1. 4. Spacecraft centered coordinate system of roll and pitch ((b and 0, respectively), and yaw (ko); all of the units are in p.rad. 5. Spacecraft centered coordinate system of roll and pitch (4) and 0, respectively), and yaw (qJ); all of the units are in #rad s- ]. 21

Original page 24

TheSimulatedSeaWiFSDataSet,Version2 O O Q r-4 _9 Cq _9 _3 ¢,q O 22

Original page 25

W.W.Gregg,F.S.Patt,andR.H.Woodward o oo" 9 Lt ¢q ¢*q O ¢4 v-4 23

Original page 26

TheSimulatedSeaWiFSDataSet,Version2 Fig. 13. A portionof Version2 simulatedSeaWiFSGACdata,band1,in satellitecoordinates(i.e., scanis x axis, orbit propagation direction is y axis). The Dominican Republic and Haiti is the dark object near the center of the image, with Puerto Rico just to the east. The grey scale indicates low at-satellite radiance (dark) to high radiance (bright). All of the bright white objects in this image are clouds. : :: 24

Original page 27

W.W.Gregg,F.S.Patt,andR.H.Woodward Fig. 14. Asin Fig.13but forband8. Landfeaturesaresaturatedin band8. 25

Original page 28

TheSimulatedSeaWiFSDataSet,Version2 .o ¢9 h_ e_ 9 C'q ¢9 f 26

Original page 29

W.W.Gregg,F.S.Patt, andR.H.Woodward visible,arethe resultof the coarseresolutionwindspeed A full moon occurs during the period of the simulated andpressurefieldsusedto computeaerosolandsunglint, data set. The required 7° phase angle for data acquisition andRayleighscatteringcontributions,respectively.Someactually occurs in the GMT morning of 27 March 1994. maybeseenin Simulation of the lunar calibration uses data collected from evidenceof minorsunglintcontamination before an actual scan of the full moon from the SeaWiFS sensor thesouthernportionofthe image,nearthecenter, of the smallsaturation in December 1992. These data produced the image shown obscurationby clouds.Because radiance in band 8 (see Table 10), this band is particularly subject to sun glint contamination, even though greater values occur at other bands. 4.2 Stored LAC Data Stored LAC data collection in the simulated data set in Simulated data acquired at the GSFC HRPT station follows the command schedule described earlier. The situ target data acquisition schedule for 25 March 1994, is shown in Fig. 15. Simulated LAC data for the second (daytime) downlink on 25 March conform to three of the in situ target sites (Fig. 16). Many of the other target sites are included in the first (nighttime) downlink on 26 March 1994 (Fig. 17), although they appear in the schedule for March 25 because actual data collection occurs at this time. The remaining targets in the 25 March schedule are downlinked at the next Wallops overpass. This scenario illustrates the difficulties in creating schedule data that occur on GMT boundaries and data collection which is tied to local noon and midnight Wallops overpasses. Actual images of the first three LAC segments in band in Fig. 22. For the simulated data set, the actual moon data were packaged into a realistic scan of two minutes in duration, showing the actual size of the moon within the downlinked SeaWiFS scan data (Fig. 23). 4.3 Real-Time HRPT Data are also provided in this data set. Mapped to Earth coordinates, a single overpass of SeaStar at the GSFC station appears as in Fig. 24, where the GSFC visibility mask is denoted. This overpass is band 1, and is the first pass of 25 March 1994. Thus, the pass may be related to the corresponding GAC image (Fig. 11). Note that the data collection does not obey the GSFC station mask boundaries. This is a result of the fact that the sensor is tilted aft of the velocity vector at this location, and the fact that the sensor scans 58 ° west-to-east of the satellite position. A non-mapped view of Version 2 simulated real-time HRPT data is provided for band 1 in Fig. 25 (also Plate 4). The Florida peninsula is visible at the western edge of the scan, with Cuba, Haiti, and the Dominican Republic, 1 show the high spatial resolution of LAC data (Fig. 18; and Puerto Rico successively west-to-east along the south- Plate 3). The first segment occurs over Saudi Arabia, and the second over South Africa, and the third over Cuba ern portion of the image. One may note significant saturation along the extreme western edge of the scans. This (Fig. 19), in a similar view of the GAC image shown ear- image corresponds directly with the GAC image shown lier (Fig. 13), but exhibiting the larger swath width of earlier (Fig. 13), as well as with the third recorded LAC LAC data. Note that the saturation at the scan edges on all three segments is due to the very large atmospheric in real-time HRPT data. path length (nearly 72 ° spacecraft zenith angle tilted) this Version 2 data. A solar calibration with an intergain segment shown in Fig. 19. This sequence of images shows the relationships among GAC data, stored LAC data, and Version 2 HRPT data for band 8 shows once again that check is also included in the LAC data on each day of the land saturates at this wavelength (Fig. 26; Plate 5). What simulated data set. A depiction of solar calibration and is particularly notable is the high resolution of land feaintergain check LAC data is shown in Fig. 20. Recall that off land mask. the procedure involved acquiring data in solar mode the solar diffuser at specified gains first, followed by intergain checks using the so-called calibration pulse. The truncated object on the left represents the solar calibration data, while the perpendicular stripe represents the results of the calibration pulse, which is solid in color until completion of the solar calibration. Then the intergain check tures, resulting from the use of the high resolution WVS 4.4 Version 2 Sensor Saturation Response The April 1993 sensor saturation response is fairly subtle and consequently is not readily apparent in the images discussed up to now. A close-up view is provided here of begins, represented by alternating bands of six scans at a portion of an HRPT image to illustrate the saturation selected gains. The TDI check is similar to the solar calibration activity, but is shorter in duration (Fig. 21). Although solar response effects. The image portion chosen is a segment on the east coast of Puerto Rico, where a number of saturating islands and coastlines occur. Band 8 was chosen because diffuser data appear brighter in the center, they are not in it saturates over these land features (bands 1-5 most likely this simulation. The increased brightness is due to different grey scaling given the absence of the intergain check. will not saturate here). The evidence is again somewhat subtle, but saturation effects appear as a kind of blurring Data from different detector combinations are assumed to preceding (stray light effects) and following (bright target be identical in this simulated set. recovery effects) small land objects (Fig. 27; Plate 6). 27

Original page 30

TheSimulatedSeaWiFSDataSet,Version2 ¢6 .-_ 0 °_ 0 0 i i ._ 28

Original page 31

W.W.Gregg,F.S.Patt,andR.H.Woodward o ¢_ o o s-, 4. (D 0 o,._ eu4 29

Original page 32

TheSimulatedSeaWiFSDataSet,Version2 ¢0 ¢fl O9 9 ._ O O ¢9 O9 4.a O9 r_ _.N < _ r..l'l 0 3O

Original page 33

W.W.Gregg,F.S.Patt, andR.H.Woodward O9 O9 4_ t_ (9 t_ .a O O9 h0 _D O9 .a 31

Original page 34

TheSimulatedSeaWiFSDataSet,Version2 .o J o O 8 O ° ..-. 32

Original page 35

W.W.Gregg,F.S.Patt, andR.H.Woodward °_ ¢D _9 cD 33

Original page 36

TheSimulatedSeaWiFSDataSet,Version2 O h_ ¢6 O O O ¢9 4_ O _6 O _9 _9 v-4 q o 34

Original page 37

W.W. Gregg, F.S. Patt, and R.H. Woodward D L.4 o o 4-) C9 o_ 35

Original page 38

The Simulated SeaWiFS Data Set, Version 2 o:Z 0._ ' . 8 _ °_ 'N -._ _ "®g o . _ _ 36

Original page 39

W.W.Gregg,F.S.Patt,andR.H.Woodward ¢9 O9 Z . O t_3 s_ t_ ¢9 Z_ 9 O9 O o 0,.-, 37

Original page 40

TheSimulatedSeaWiFSDataSet,Version2 t_ ,4 CD 9 4 o .o 4_ o _J o5 38

Original page 41

W.W. Gregg, F.S. Patt, and R.H. Woodward cD w 8_ W 39

Original page 42

The Simulated SeaWiFS 5. DATA AVAILABILITY The data set containing 16-bit science data (frame formatter output or level-0 data) is available on Internet from the GSFC DAAC. The address is: eosdata.gsfc.nasa.gov. Log in as anonymous with password guest, then change directories to pub/seawifs. All files begin with S followed by the year, so all files belonging to this simulated data set (Version 2) will begin with S1994. GAC files end with the suffix LO_GAC, recorded LAC files with LO_LAC and HRPT files with L0_HRPT. Assistance may be obtained by dialing 301-286-3209. Please note that for the mission, authorization must be obtained before acquiring the data. The data files containing 10-bit science data (onboard structure) were intended to be used internally for endto-end system testing. The data are available from the authors on 4mm tape for Silicon Graphics IRIX or Sun SunOS systems. ACKNOWLEDGMENTS The authors are indebted to a large number of people for their assistance and advice which was required due to the complexity of this undertaking. They are in alphabetical order: Robert Barnes (Man Tech Environmental), Frank Corprew (Hughes- STX), Gene Feldman (GSFC), Mary James (GSFC), Ken Lambert (University of Maryland), Robert Mack (Hughes-STX), and George Riggs (Research and Data Systems, Corp.). We would also like to thank Wayne Esaias for critical comments on the manuscript, and for encouraging us to submit it despite the sensor modifications agreed upon after the completion of this effort. GLOSSARY AMC Angular Momentum Compensation AVHRR Advanced Very High Resolution Radiometer BTR Bright Target Recovery CZCS Coastal Zone Color Scanner DAAC Distributed Active Archive Center DCF Data Capture Facility ECEF Earth-Centered Earth-Fixed FNOC Fleet Numerical Oceanography Center FPA Focal Point Assembly GAC Global Area Coverage GMT Greenwich Mean Time GPS Global Positioning System GSFC Goddard Space Flight Center HRPT High Resolution Picture Transmission ICD Interface Control Document ISCCP International Satellite Cloud Climatology Project LAC Local Area Coverage LSB Least Significant Bits MSB Most Significant Bits NASA National Aeronautics and Space Administration NASCOM NASA Communications NCDS NASA Climate Data System NDVI Normalized Difference Vegetation Index NER Noise Equivalent Radiance NGDC National Geophysical Data Center 40 Data Set, Version 2 NIMBUS Not an acronym, a series of NASA experimental weather satellites containing a wide variety of atmosphere, ice, and ocean sensors. North American Air Defense (Command) NORAD OSC Orbital Sciences Corporation PFF Programmable Frame Formatter QC Quality Control SBRC (Hughes) Santa Barbara Research Center S/C Spacecraft SeaWiFS Data Processing System SDPS Sea-viewing Wide Field-of-view Sensor SeaWiFS SOC Simulation Operations Center SOH State of Health TDI Time-Delay and Integration TOMS Total Ozone Mapping Spectrometer WFF Wallops Flight Facility World Vector Shoreline WVS SYMBOLS a A constant equal to -20/tanh(2). ao Oxygen absorption coefficient. aoz Ozone absorption coefficient. awv Coefficient for water vapor absorption. b A constant equal to 1/3. DC Digital count value. Ed Incident downwelling irradiance. Fo(A) Mean extraterrestrial spectral irradiance. G Gain factor. LNERA) Noise equivalent radiance. L_.() Saturation radiance for the sensor. L_(A) At-satellite radiance. M Atmospheric slant path length. P Local surface pressure. P0 Standard pressure. 8(_) Slope for the range 0-1,023. t Time in seconds. T Tilt position. V Volts. X ECEF X component of orbit position. 2 ECEF X component of orbit velocity. Y ECEF Y component of orbit position. ECEF Y component of orbit velocity. Z ECEF Z component of orbit position. 2 ECEF Z component of orbit velocity. Wavelength of light. ql Yaw. Yaw rate. Roll. Roll rate. pc,i Reflectance of clouds and ice. 0 Pitch. Pitch rate. GoSolar zenith angle. Tr Rayleigh optical thickness. r, Pressure corrected Rayleigh optical thickness. Tro Rayleigh optical thickness weighted by the SeaWiFS spectral response.

Original page 43

W.W.Gregg,F.S.Patt,andR.H.Woodward REFERENCES Gordon, H.R., J.W. Brown, and R.H. Evans, 1988: Exact Rayleigh scattering calculations for use with the Nimbus-7 Coastal Zone Color Scanner. Appl. Opt., 862-871. Gregg, W.W., and K.L. Carder, 1990: A simple spectral solar irradiance model for cloudless maritime atmospheres. Limnol. Ocean., 1,657-1,675. , F. Chen, A. Mezaache, J. Chen, and J. Whiting, 1993: The simulated SeaWiFS data set: Version 1. NASA Teeh. Memo. 104566, Vol. 9, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 17pp. ---, and F.S. Patt, 1993: Assessment of tilt capability for spaceborne global ocean color sensors. IEEE Trans. Geosci. Remote Sens., (submitted). Patt, F.S., and W.W. Gregg, 1993: Exact closed-form geolocation algorithm for Earth survey sensors. Inter. J. Remote Sens., (submitted). Tucker, C.J., and L.D. Miller, 1977: Soil spectra contributions to grass canopy spectral reflectance. Photogrammetry, Engineering, and Remote Sens., 721-726. Wertz, J.R. (Ed.), 1978: Spacecraft Attitude Determination and Control. D. Reidel, Dordrecht, Holland, 858 pp. Williams, S.P., E.F. Szajna, and W.A. Hods, 1985: Nimbus 7 Coastal Zone Color Scanner (CZCS) level-1 data product users' guide. NASA Tech. Memo. 86_03, NASA Goddard Space Flight Center, Greenbelt, Maryland, 49 pp. Woodward, R.H., R.A. Barnes, C.R. McClain, W.E. Esalas, W.L. Barnes, and A.T. Mecherikunnel, 1993: Modeling of the SeaWiFS solar and lunar observations. NASA Tech. Memo. 104566, Vol. 10, S.B. Hooker and E.R. Firestone, Eds., 26 pp. THE SEAWIFS TECHNICAL REPORT SERIES Vo/./ Hooker, S.B., W.E. Esaias, G.C. Feldman, W.W. Gregg, and C.R. McClain, 1992: An Overview of SeaWiFS and Ocean Color. NASA Tech. Memo. 104566, Vol. 1, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 24pp., plus color plates. Yo/.2 Gregg, W.W., 1992: Analysis of Orbit Selection for SeaWiFS: Ascending vs. Descending Node. NASA Tech. Memo. 104566, Vol. g, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 16 pp. VoI.3 McClaln, C.R., W.E. Esaias, W. Barnes, B. Guenther, D. Endres, S. Hooker, G. Mitchell, and R. Barnes, 1992: Calibration and Validation Plan for SeaWiFS. NASA Tech. Memo. 104566, Vol. 3, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 41 pp. VoL 4 McClain, C.R., E. Yeh, and G. Fu, 1992: An Analysis of GAC Sampling Algorithms: A Case Study. NASA Tech. Memo. 104566, VoL 4, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 22 pp., plus color plates. Mueller, J.L., and R.W. Austin, 1992: Ocean Optics Protocols. NASA Tech. Memo. 104566, Vol. 5, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 43 pp. Vol. 6 Firestone, E.R., and S.B. Hooker, 1992: SeaWiFS Technical Report Series Summary Index: Volumes 1-5. NASA Tech. Memo. 104566, Vol. 6, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 9 pp. Voi. 7 Darzi, M., 1992: Cloud Screening for Polar Orbiting Visible and IR Satellite Sensors. NASA Tech. Memo. 104566, Vol. 7, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 7pp. Hooker, S.B., W.E. Esaias, and L.A. Rexrode, 1993: Proceedings of the First SeaWiFS Science Team Meeting. NASA Tech. Memo. 104566, Vol. 8, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 61 pp. Yol. 9 Gregg, W.W., F.C. Chen, A.L. Mezaache, J.D. Chen, J.A. Whiting, 1993: The Simulated SeaWiFS Data Set, Version 1. NASA Tech. Memo. 104566, Vol. 9, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 17 pp. yo/./0 Woodward, R.H., R.A. Barnes, C.R. MeClain, W.E. Esaias, W.L. Barnes, and A.T. Mecherikunnel, 1993: Modeling of the SeaWiFS Solar and Lunar Observations. NASA Tech. Memo. 104566, Vol. 10, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 26pp. _Vol. 1/ Patt, F.S., C.M. Hoisington, W.W. Gregg, and P.L. Coronado, 1993: Analysis of Selected Orbit Propagation Models for the SeaWiFS Mission. NASA Tech. Memo. 104566, Vol. 11, S.B. Hooker, E.R. Firestone, and A.W. Indest, Eds., NASA Goddard Space Plight Center, Greenbelt, Maryland, 16 pp. yol. 12 Firestone, E.R., and S.B. Hooker, 1993: SeaWiFS Technical Report Series Summary Index: Volumes 1-11. NASA Tech. Memo. 104566, Vol. 12, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 28 pp. 41

Original page 44

TheSimulatedSeaWiFSDataSet,Version2 Vol. 13 McClain, C.R., K.R. Arrigo, J. Comiso, R. Fraser, M. Darzi, J.K. Firestone, B. Schieber, E-n. Yeh, and C.W. Sullivan, 1993: Case Studies for SeaWiFS Calibration and Validation, Part 1. NASA Tech. Memo. 10._566, Vol. 13, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 52pp., plus color plates. Vol. 14 Mueller, J.L., 1993: The First SeaWiFS Intercalibration Round- Robin Experiment, SIRREX-1, July 1992. NASA Tech. Memo. 104566, Vol. 14, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 60pp. 42 Vol. 15 Gregg, W.W., F.S. Patt, R.H. Woodward, 1993: The Simulated SeaWiFS Data Set, Version 2. NASA Tech. Memo. i0_566, Vol. 15, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 42pp., plus color plates.

Original page 45

W.W. Gregg, F.S. Patt, and R.H. Woodward COLOR PLATES The following color plates are presented as submitted by the authors.

Original page 46

TheSimulatedSeaWiFSDataSet,Version2 PLATE1. Top:SimulatedSeaWiFSGACdata,25March1994,band1,mappedto Earthcoordinates. Bottom:band8.

Original page 47

W.W.Gregg,F.S.Patt,andR.H.Woodward PLATE2. Left: A portionofsimulatedSeaWiFSGACdata,band1,insatellitecoordinates(i.e.,scanis x axis, orbit propagation direction is y axis). The Dominican Republic and Haiti is the dark blue object near the center of the image, with Puerto Rico to the east. The color scale indicates low at-satellite radiance (dark blue) to high radiance (bright red). All of the bright red objects in this image are clouds. Left: band 8. Land features are saturated in band 8.

Original page 48

TheSimulatedSeaWiFSDataSet,Version2 a9 i .4 ¢6 5O qJ aS °_ O | O ¢,9 5 .= qJ . al,-_

Original page 49

W.W. Gregg, F.S. Patt, and R.H. Woodward o ct_ 8_ o o 8 °,._ =_ N)

Original page 50

The Simulated SeaWiFS Data Set, Version 2 e_ o e_ o ..q e _o

Original page 51

The Simulated SeaWiFS Data Set, Version 2 u r ,,,,,,,I __

Original page 52

i Form Approved REPORT DOCUMENTATION PAGE OMeNo.o7o,-o1 Public repoding burden for this collection of information1 is estimated to average 1 hour per response, including the time lor reviewing instructJon,L searching existing data sources, oathering and maintaining the data needed, and completing and reviewing the collection of infoanation. ,)nd comrnlmts regarding this burden estimate _ any other aspect of this collection of informatlo, inciuding sugge6tio_s fo¢ reducing this burden, to Washington Heedquarters Se_ices, Directorate f_ Info(mation Operations and Reports, 1215 Jefferson [:)avis Highway, Suite 1204 rAdington= VA 99909-43021 and to the Office of Management and Budget r Pat:lerwork Reduction Pro_ect 0704-0188} WashM_lon t DC 20503. 1. AGENCYUSEONLY (Leaveb/_k) 2. REPORTDATE January 1994 4. TITLE ANDSUBTITLE SeaWiFS Technical Report Series Volume 15-The Simulated SeaWiFS Data Set, Version 2 6. AUTHOR(S) Watson W. Gregg, Frederick S. Patt, and Robert H. Woodward Series Editors: Stanford B. Hooker and Elaine R. Firestone 7. PERFORMINGORGANIZATIONNAME(S)AND ADDRESS{ES) Laboratory for Hydrospheric Processes Goddard Space Flight Center Greenbelt, Maryland 20771 3. REPORTTYPE ANDDATESCOVEREb Technical Memorandum 5. FUNDINGNUMBERS 8. PERFORMINGORGANIZATION REPORTNUMBER 94B00041 Code 970.2 ANDADORESS(ES) 10. SPONSORING/MONITORING 9. SPONSORING/MONITORINGAGENCYNAME(S) National Aeronautics and Space Administration Washington, D.C. 20546--0001 11. SUPPLEMENTARYNOTES AGENCYREPORTNUMBER TM-104566, Vol. 15 Frederick S. Patt, Robert H. Woodward, and Elaine R. Firestone: General Sciences Corporation, Laurel, Maryland. 12a. DISTRIBUllON/AVAILABlUTY STATEMENT Unclassified-Unlimited Subject Category 48 Report is available from the National Technical Information 12b. DISTRIBUTIONCODE Service, U.S. Dept. of Commerce, 5285 Port Royal Road, Springfield, VA 22151; (703) 557-4650. 13. ABSTRACT(Maximum200 words) This docuraent describes the second version of the simulated SeaWiFS data set. A realistic simulated data set is essential for mission readiness preparations and can potentially assist in all phases of ground support for a future mission. The second version improves on the first version primarily through additional realism and complexity. This version incorporates a representation of virtually every aspect of the flight mission. Thus, it provides a high-fidelity data set for testing several aspects of the ground system, including data acquisition, data processing, data transfers, calibration and validation, quality control, and mission operations. The data set is constructed for a seven-day period, 25-31 March 1994. Specific features of the data set ir_ude Global Area Coverage (GAC), recorded Local Area Coverage (LAC), and reallime High Resolution Picture Transmission (HRIrl ") data for the seven-day period. A realistic orbit, which is propagated using a Brouwer-Lyddane model with drag, is used to simulate orbit positions° The simulated data corresponds to the command schedule based on the orbit for this seven-day period. It includes total (at-satellite) radiances not only for ocean, but for land, clouds, and ice. The simulation also utilizes a high-resolution land-sea mask. It includes the April 1993 SeaWiFS spectral responses and sensor saturation responses. The simulation is formatted according to July 1993 onboard data sb'uctures, which include corresponding telemetry (instnmaent and spacecraft) data The methods are described and some examples of the output are given. The instrument response functions made available in April 1993 have been used to produce the Version 2 simulated data= These response functions will change as part of the sensor improvements initiated in July-August 1993. 14. SUBJECTTERMS Propagation, Navigation, 51 SeaWiFS, Oceanography, GAC Data, LAC Data, Orbit 15. NUMBEROF PAGES Command Schedules, At-Satellite Radiances, Sensor Saturation Response, 16o PRICE CODE Data Formatting, HRPT Data 17.SECURITYCLASSIFICATION 18.SECURITYCLASSIFICATION OF REPORT OF THIS PAGE Unclassified Unclassified NSN 7540-01-280-5500 19.SECURITYCLASSIFICATION 20. UlVlITATIONOF AS,%q'RAC-T OF ABSTRACT Unclassified Unlimited Standard Form 298 (Rev. 2-89) Prel_rlbed by ANSI S'lXL 239-18, 298-102